A man who signed guarantees worth around ₹22,000 crore will pay ₹6.5 crore, and a tribunal has said that is lawful. That sentence is accurate, and on its own it is close to useless. It tells you nothing about who the money is owed to, why nobody could get more, which creditors were outvoted, or what happens to the loans themselves once this proceeding closes.

The order approving the plan runs to well over a hundred pages, and the reporting around it has been compressed into a single percentage. This piece goes the other way. It works through the creditor list line by line, recomputes the vote from the schedule, converts the recovery rate into rupees for every lender in the case, and separates what the tribunal actually decided from what commentators have assumed it decided.

Everything below is attributed. Where a figure is our own arithmetic rather than a reported number, it says so. Where reports disagree, both versions appear.

It started with one loan, not with ₹22,000 crore

The scale of the final number hides how small the trigger was. Bar and Bench reports that Indiabulls Housing Finance, now Sammaan Capital, moved the tribunal in 2022 under Section 95 of the Insolvency and Bankruptcy Code after Vivek Infracon defaulted on a ₹170 crore loan on which Chandra had signed a personal guarantee.

Section 95 is the provision that lets a creditor start insolvency proceedings against a personal guarantor. Once the plea was admitted in 2024, the process stopped being about one loan. A resolution professional was appointed, claims were invited, and every creditor holding a guarantee signed by the same individual could file. The Week reports that 23 creditors came forward and the admitted total climbed to ₹22,006.57 crore.

That is the mechanism that turned ₹170 crore into ₹22,006.57 crore. Not new borrowing, not fresh default, but the aggregation of every guarantee one person had signed for a group of companies into a single personal estate proceeding.

Debt, claim and recovery are three separate quantities

Almost every misreading of this order comes from collapsing three different things into one word. Getting them apart is the whole exercise.

Business Standard states the position on this case plainly: the ₹22,006.57 crore represents claims admitted against Chandra in his capacity as guarantor for loans taken by companies and other principal debtors, and a guarantor’s liability can be far larger than the assets he personally owns. The principal borrowers remain separate entities with their own obligations.

So “99.97 per cent of the debt has been written off” is not what happened. What happened is that 99.97 per cent of the claims filed against one individual’s personal estate will go unpaid from that estate. The loans themselves sit elsewhere, with different obligors and, in several cases, with security still attached.

The full creditor roll call, with what each one stands to get

Voting share in a personal insolvency proceeding is proportional to admitted claim value, which means the schedule of creditors is also, in effect, the schedule of who decides the outcome. Here is the complete list of 23, with the stance each took on the repayment plan and the pro-rata payout their voting share implies from the ₹6.25 crore corpus.

The payout column is our calculation, not a published schedule. It applies each creditor’s voting share to the ₹6.25 crore corpus. The method checks out against the one figure the order does record: LIC Housing Finance’s 6.09 per cent of ₹6.25 crore is ₹38.06 lakh, and the order records ₹38.09 lakh. Business Standard has reported that the approval is subject to changes in the list of eligible creditors and redistribution of the corpus, so treat every other row as an illustration of scale rather than a final entitlement.

Read down that column and the practical meaning of a 99.97 per cent haircut becomes concrete. The single largest creditor in the case, holding more than a quarter of all admitted claims, receives under ₹1.8 crore. Three of the twenty-three receive less than ₹1.2 lakh each. These are amounts a mid-sized company would treat as a rounding error in a quarterly provision.

The vote, recomputed from the schedule

The 80.81 per cent approval figure did most of the heavy lifting in the tribunal’s reasoning, so it deserves to be taken apart rather than repeated. Aggregating the schedule gives three blocs: ten creditors holding 77.48 per cent voted in favour, seven holding 18.39 per cent voted against, and six holding 4.10 per cent did not vote at all.

World Crest Advisors at 28.49 per cent, Lemonade Capital Advisors at 16.85 per cent, Catalyst Trusteeship acting for a CINDA fund at 11.85 per cent and Corpcall Capital Advisors at 10.30 per cent together held 67.49 per cent of the vote. All four supported the plan. The seven institutions that opposed it held 18.39 per cent between them. On voting weight, the banks were never in a position to block anything.

This is not a technicality. A credit officer reading “80.81 per cent of creditors approved” will assume that four in every five rupees of claim actively supported the outcome. In fact, just over 77 in every 100 did. Abstention in a voting-share regime is not neutrality; it is silent support for whichever side is ahead.

What each objecting lender actually receives

The seven institutions that voted against the plan hold ₹4,047 crore of claims between them on our calculation from their voting shares. Their combined entitlement under the plan is about ₹1.15 crore. The chart below breaks that down lender by lender.

LIC Housing Finance told the tribunal the payout was unviable and unlawful, and Bar and Bench reports it also argued that even the ₹6.5 crore was described in the plan as indicative and uncertain. A senior banking executive quoted by Business Standard was blunter, saying a plan with this much haircut is totally unviable and that an NCLAT appeal was being explored.

One conversion makes the point stick. At 0.0284 per cent, a creditor receives about ₹2,840 for every ₹1 crore of admitted claim, or roughly 28 paise for every ₹1,000. LIC Housing Finance, on the recorded figure, does marginally better at about 29 paise per ₹1,000.

Why the tribunal refused to price the claims itself

The obvious question is why an adjudicating authority accepted a number that every institutional lender in the room called indefensible. The answer sits in what Section 114 actually asks a tribunal to do.

Business Standard’s explainer sets out the position: the IBC does not fix a universal minimum percentage that creditors must recover under a personal-guarantor repayment plan. It prescribes a process instead. The guarantor prepares a plan with the resolution professional, creditors consider and vote on it, and the adjudicating authority then examines whether the plan can be approved under the Code.

The tribunal said it could not substitute its own commercial judgment for that of the creditors. Its questions were whether the statutory process had been followed, whether the voting was valid, whether the plan complied with the Code, and whether there was any legal basis to interfere. Not whether ₹6.25 crore was a fair price for ₹22,006.57 crore of claims.

Whether that valuation is right is a separate question from whether the tribunal was entitled to rely on it, and it is the question an appellate bench would have to engage with.

Sections 115 and 138: the part dissenting lenders cannot walk away from

A creditor who votes against a plan might reasonably expect to keep its own rights intact. Under the Code, it does not.

Business Standard reports that Section 115 of the IBC governs the effect of an approved repayment plan, and that the tribunal held the plan binds creditors covered by it, including those who voted against. Allowing dissenters to pursue their full original claims outside the plan would, in the tribunal’s reasoning, defeat the collective character of insolvency. The order also notes that once a discharge order is passed under Section 138, creditors cannot continue to pursue the personal guarantor for past dues covered by that discharge.

The practical consequence is worth stating flatly. HDFC Bank, Axis Bank, Canara Bank, RBL Bank, Union Bank of India (UK), LIC Housing Finance and IDBI Trusteeship for the Franklin Templeton fund all voted no, and all of them are inside the plan. Their remedy is an appeal, not a private enforcement route against the guarantor.

From ₹45,888 crore to ₹31.79 crore: the gap the order left open

The objecting creditors’ strongest point was never the percentage. It was the estate. Bar and Bench reports that they relied on net-worth certificates showing Chandra’s net worth at approximately ₹45,888 crore in 2017 and ₹40,562 crore in 2018, against a present stated net worth of about ₹31.79 crore.

What the third member did with that gap is the finding most likely to be tested on appeal. Business Standard’s account is careful: the order records the objections based on those historical figures, but does not provide a complete, asset-by-asset explanation of how the wealth declined to the current valuation. The third member held that the discrepancy alone did not establish concealment, and did not make a forensic investigation mandatory before the plan could be considered.

That is a ruling about the evidentiary threshold for ordering an investigation. It is not a finding that the decline has been explained, and the order does not claim to be one.

The associate-vote question that could have changed the result

There was a second line of attack, and on the arithmetic it was the more dangerous one. Business Standard reports that dissenting creditors argued some entities supporting the plan were associated with Chandra or the Essel Group. That matters because creditors falling within the statutory definition of an associate can face restrictions on voting.

Run the numbers on what that would have meant. If entities holding, say, the largest single supporting position had been excluded from the count, the supporting bloc would have fallen from 77.48 per cent to under 50 per cent of total voting share, and the plan would not have carried. The margin between approval and rejection was never about the banks. It was about whether four large non-bank holders were entitled to vote.

The third member found the evidence insufficient to establish a legal basis for excluding those votes, so they remained part of the approving majority. That is a finding on sufficiency of evidence rather than on the underlying relationships, which is precisely the kind of finding appellate benches are asked to revisit.

The ₹1,494 crore that never made the headline

Here is the number that changes the meaning of the story, and it appeared in almost no coverage. Business Standard reports that government sources sought to correct the impression that the order meant a 99.97 per cent loss on the underlying loans, and said the repayment plan envisages about ₹1,494 crore in payments by the principal corporate borrowers, separate from Chandra’s personal contribution.

That is roughly 239 times the ₹6.25 crore corpus. It is not money received, it is not enforced by this order, and it has not been published as a schedule with dates or security attached. But it is the reason the personal insolvency figure and the eventual loan-loss figure are not the same figure.

Reporting carried by Storyboard18, citing the Economic Times, adds a related point from LIC Housing Finance itself: the lender has said the NCLT order does not affect the underlying liabilities of the corporate borrowers or its security. A dissenting creditor bound by the plan against the guarantor is not bound in its dealings with the borrower.

What Subhash Chandra says the numbers actually are

On 27 August 2026, Chandra issued a statement disputing the reporting. He said there is no personal borrowing by him from any creditor named in the order and that he only signed personal guarantees, and he urged media organisations to correct information he described as incorrectly interpreted from the court order.

His central factual claim is that the relevant claims total ₹3,992 crore rather than ₹22,000 crore. He said ₹620 crore of that has already been settled, and that lenders whose claims totalled ₹16,386 crore had either accepted the plan or filed no objection, with admitted claims for that group at ₹16,201 crore. He also said the borrowing entities have repaid approximately ₹43,000 crore since January 2019, against total outstanding borrowings of about ₹45,000 crore as of 24 January 2019.

One observation follows from putting the two sets of numbers side by side, and it is ours rather than anyone else’s. The seven objecting institutions held 18.39 per cent of voting share, which maps to roughly ₹4,047 crore of admitted claims. That is close to the ₹3,992 crore Chandra calls the relevant claim total. The two sides may be describing the same objecting bloc and disagreeing only about whether the non-objecting claims should be counted in the headline at all.

How rare is an approved personal-guarantor plan in India?

Rare enough that insolvency practitioners are reading one order about one businessman with unusual attention. Personal guarantor insolvency is the youngest and thinnest part of the Code: the provisions came into force only on 1 December 2019.

The regulator has been closing exactly the gaps this case exposed. Amendments effective 2 June 2026 inserted a new Regulation 6A requiring a comprehensive asset statement from the guarantor, replaced Regulation 11A to coordinate asset transfers, and tightened Regulation 17B, which deals with cases admitted where no repayment plan is ever filed. A guarantor admitted into the process after that date faces a materially heavier disclosure burden than one admitted in 2024, which is when this case was admitted.

Where a promoter guarantee really sits on the recovery scale

For a credit committee, the useful output of this order is a repricing question rather than a moral one. What is an unsecured promoter guarantee worth once the personal estate has been depleted? The rail below places recovery outcomes into bands and states what each implies for how the instrument should be carried.

The comparison is not like for like, because Videocon was a corporate resolution with operating assets while this is an individual’s estate. It is still the right order-of-magnitude reference, and it explains why practitioners describe this outcome as a category of its own rather than a bad result within a familiar range.

Four routes that survive the plan

The order closes one door and leaves several open. Ranked by how directly each bears on actual recovery, here is what remains available to a lender bound by the plan.

Eight things this order changes about underwriting a promoter guarantee

None of this is legal advice. It is the practical reading a credit committee would take from the order and the reporting around it.

The vocabulary of the order, decoded

Several terms in this case were used loosely in the coverage, and the looseness is where the misunderstanding lives. This table gives each one in plain language and says why it matters here.

What we know, and what is still unclear

Separating the confirmed from the unconfirmed is the most useful thing an explainer can do while a matter is still live. The first list is recorded in the order or in reporting on it. The second is not.

What we know

• NCLT Member (Judicial) Nilesh Sharma, sitting as a third member on the Delhi bench after a split verdict, approved the repayment plan on 25 August 2026 under Section 114 of the IBC.
• The plan provides ₹6.25 crore for creditors and ₹25 lakh for process costs against admitted claims of ₹22,006.57 crore, a haircut of about 99.97 per cent.
• Creditors holding 80.81 per cent of voting share approved it; the schedule shows ten creditors in favour, seven against and six not voting.
• LIC Housing Finance had an admitted claim of ₹1,322.39 crore and the order records a proposed payment of ₹38.09 lakh, which the lender called unviable and unlawful.
• The case began with Indiabulls Housing Finance, now Sammaan Capital, moving under Section 95 in 2022 over a ₹170 crore loan to Vivek Infracon.
• The original split was between Judicial Member Ashok Kumar Bhardwaj and Technical Member Reena Sinha Puri, and the third member was appointed under Section 419(5) of the Companies Act, 2013.
• HDFC Bank has said it is exploring an NCLAT appeal, and LIC Housing Finance has been reported as deciding to challenge the order and to explore intervention through the National Housing Bank.

What is still unclear

• The date on which creditors voted on the repayment plan has not been confirmed in the reporting reviewed for this article.
• The final amount each creditor receives is not settled, because the approval is subject to changes in the eligible creditor list and redistribution of the corpus.
• Whether the signalled appeals have been filed, on what grounds, and whether any stay has been sought had not been confirmed as at 28 August 2026.
• How the ₹1,494 crore expected from principal borrowers is structured, secured or timed has not been published in the reporting reviewed here.
• The asset-by-asset explanation for the fall in stated net worth from ₹45,888 crore in 2017 to about ₹31.79 crore is not set out in the order as reported.
• Whether Chandra’s ₹3,992 crore figure and the ₹22,006.57 crore admitted total can be reconciled, and on what basis, has not been explained by either side.
• The identity and relationships of the four largest supporting creditors have not been detailed publicly beyond the allegation that was rejected for want of evidence.

Frequently asked questions

The short version

A tribunal approved a plan under which one man pays ₹6.5 crore against ₹22,006.57 crore of claims admitted against him as a guarantor. It did so because creditors holding 80.81 per cent of votes cast said yes, because the Code sets no minimum recovery, and because the valuation found his personal estate was worth less than the plan offered. Four non-bank entities carried 67.49 per cent of the vote. The seven institutions that objected held 18.39 per cent, will share about ₹1.15 crore between them, and are bound by the plan regardless.

What the order does not do is erase the underlying loans. The borrowers remain liable, security remains enforceable, about ₹1,494 crore is expected from those companies, and two lenders are heading to the appellate tribunal. The number to watch next is not 99.97 per cent. It is what the NCLAT does with the associate-vote finding, and how much of that ₹1,494 crore actually arrives.